Introduction Low initial differential renal function (DRF) in patients with primary non-refluxing megaureter (PNRM) is considered an indication for surgery as are an increase of dilatation and symptoms. Objective We hypothesized that low DRF is not necessarily a result of obstruction, but may be due to impaired development of the upper urinary tract. Thus, in the absence of symptoms, there is a low risk for further loss of renal function. This study aimed to assess whether initially low DRF is a reliable indicator of obstruction. Study design We reviewed data from four university centers between 1995 and 2010. Patients under 12 months of age with unilateral primary non-refluxing megaureter (PNMR) and a DRF between 10% and 40%, and followed minimally 24 months, were included. Patients were placed in two groups based on management: group A, surgical; group B, conservative. The dynamics of DRF in relation to age and type of treatment was studied. In each patient we recorded the earliest (initial) DRF, the last known (final) DRF, the age when MAG-3 scans were performed and the type of treatment. Results From 25 patients, 16 were treated surgically (group A) and 9 followed conservatively (group B). The initial mean DRF in group A was 33.1% and in group B 34.5%, at a mean age 3.0 and 3.6 months, respectively. The final mean DRF in group A was 40.1% and in group B 43%, at a mean age 59.9 and 46.3 months, respectively. Using two-way repeated ANOVA (age [initial DRF, final DRF] vs. group [group A, group B]), we found non-significant difference between the groups in the DRF, F (1, 21) = 0.96, p = 0.338, while we observed statistically significant and similar increase from the initial to final DRF in both groups, F (1, 21) = 16.66, p = 0.001 (Figure). Discussion This is the first study focusing on the evolution of renal function in patients with PNRM and low initial DRF. Results suggest that the diagnosis of obstruction is inaccurate in most infants with unilateral PNRM if it is based on low initial DRF only. Renal deterioration rarely occurs in asymptomatic patients, and even profoundly impaired kidneys have potential for improvement. Limitations of our study include retrospective design and lack of standardization of treatment among the four centers. Conclusion Low DRF in asymptomatic and anatomically stable patients with PNMR should not be considered an indication for early surgery. These findings challenge current practice and should be confirmed by a prospective study.